Software Factory vs DevOps: When Agents Do the Work

By Paula Hingel (Augment Code).

Core Shift

DevOps organized handoffs between human developers and operations. A software factory changes who performs the labor inside the pipeline: AI agents handle production work from intake to ship, while humans supervise at defined exception points.

Bottleneck Shift

In human-led delivery, the primary constraint is authoring speed. When agents write code, the bottleneck flips to downstream verification:

Specification Quality: Ambiguity directly drives rework and hallucinations.
Execution Context: Context engines and permission-aware retrieval prevent stale assumptions.
Verification: Risk-tiered gates addressing non-deterministic agent failures (syntactically plausible code with semantic bugs).
Traceability: Immutable audit trails for agent-generated diffs and prompt histories.
Controlled Iteration: Bounded blast radius and vertical-slice decomposition.

Risk-Tiered Human Oversight Gates

Low-risk changes: automated checks only (lint, types, deterministic unit tests).
Standard work: single-reviewer gate / automated adversarial review.
High-blast-radius changes: multi-reviewer gate and mandatory human sign-off on destructive actions (production deploys, migrations, auth boundaries).